hi linuxers,

i spent the day trying to move my system (RH9) from an old IDE drive to a 9.1 GB SCSI drive. first i tried using DUMP and RESTORE to copy the relavent partitions [ boot / ] from [ hda1 hda2 ] to [ sda1 sda2 ]. after setting the sda1 boot flag, rebooting to the SCSI drive resulted in silent black screen.

not knowing enough about kernels and grub, i tried to fix the non-booting condition by loading a minimal install of redhat on the SCSI drive with the same partition scheme as the working IDE setup ( [ sda1 sda2 sda3 ] =~ [ hda1 hda2 hda3 ] = [ boot / swap ] ). Then RESTORE was used to recreate the working / structure from hda2 on sda2. This resulted in a kernel panic upon rebooting to the SCSI drive, with linux complaining about "no init found, try passing init= " to the kernel.

then RESTORE was used to move the previously working minimal install / partition back to sda2 and the same kernel panic was returned upon reboot.

thus two questions remain:
1. does anybody have a suggestion how to successfully move my working IDE setup onto a SCSI drive?
2. the last time RESTORE was used, it copied an image to sda2 that worked on sda2 when the DUMP was created. this shows my use of DUMP/RESTORE is not a viable means to backup my system. can anybody suggest why not? the commands i chose to use were 'dump -0f /mnt/Files/sda2.dump /dev/sda2' and 'restore rf /mnt/Files/sda2.dump'.

thanks if any ideas. -brian

The problem probably lies in your fstab file. You need to edit that and change all instances of /dev/hdaX to /dev/sdaX. Also you will need to edit /etc/lilo.conf and do the same thing, then save lilo.conf and run lilo.

thx for the useful suggestions. i'll try that now.

i chose grub not lilo when reloading the basic redhat install onto the SCSI drive, if i can't work your suggestions into grub i'll prob just reload with lilo and try your ideas.

anybody have a quick pointer on which files to edit to get grub to jive with sda instead of hda? i never had to play with grub/lilo setup stuff before.

also, does anybody have a guess why dumping sda then restoring the same sda an hour later gives a kernel panic also? i was hoping to use dump/restore as a long-term backup method but now question my method of implementingf it.

thanks for taking the time to read, - brian

Well besides the /etc/fstab problems you also have the root dev setting in the kernel. Lilo passes the kernel root=/dev/hda2 but if / is on /dev/sda2 then the kernel can't find /sbin/init. It's possible that the RedHat kernel didn't have scsi drivers built into it. Maybe they were built as modules and therefore the kernel needs an initrd file.

I think your idea for backup will work as long as you return the files back to the original type of drive.

i changed all references in grub.conf and fstab from hdaX to sdaX.

when trying to boot, grub reports:
Error 20: Error while parsing number
i am unsure how to benefit from this message.

thx for the new advice. SCSI support is there since a burner always sits on the narrow bus. i also tried changing the grub.conf line
#boot=/dev/hda
so grub.conf looks like this:

boot=/dev/sda
default=1
timeout=3
splashimage=(sd0,0)/grub/splash.xpm/gz
title Red Hat Linux (2.4.20-8)
root (sd0,0)
kernel /vmlinuz-2.4.20-8 ro root=LABEL=/
initrd /initrd-2.4.20-8.img

any additional comments are appreciated. i'll try lilo next weekend but guess that the problem lies in the kernel somewhere, outside of my immediate reach.

thx for reading, -brian